More Like this

1. CAVE: Connectome Annotation Versioning Engine

   https://doi.org/10.1038/s41592-024-02426-z  

   Dorkenwald, Sven; Schneider-Mizell, Casey_M; Brittain, Derrick; Halageri, Akhilesh; Jordan, Chris; Kemnitz, Nico; Castro, Manual_A; Silversmith, William; Maitin-Shephard, Jeremy; Troidl, Jakob; et al (April 2025, Nature Methods)

   Abstract Advances in electron microscopy, image segmentation and computational infrastructure have given rise to large-scale and richly annotated connectomic datasets, which are increasingly shared across communities. To enable collaboration, users need to be able to concurrently create annotations and correct errors in the automated segmentation by proofreading. In large datasets, every proofreading edit relabels cell identities of millions of voxels and thousands of annotations like synapses. For analysis, users require immediate and reproducible access to this changing and expanding data landscape. Here we present the Connectome Annotation Versioning Engine (CAVE), a computational infrastructure that provides scalable solutions for proofreading and flexible annotation support for fast analysis queries at arbitrary time points. Deployed as a suite of web services, CAVE empowers distributed communities to perform reproducible connectome analysis in up to petascale datasets (~1 mm³) while proofreading and annotating is ongoing. 

   more » « less
2. VICE: Visual Identification and Correction of Neural Circuit Errors

   https://doi.org/10.1111/cgf.14320  

   Gonda, F; Wang, S; Beyer, J; Lichtman, J; Pfister, H (January 2021, Eurographics Conference on Visualization (EuroVis))

   A connectivity graph of neurons at the resolution of single synapses provides scientists with a tool for understanding the nervous system in health and disease. Recent advances in automatic image segmentation and synapse prediction in electron microscopy (EM) datasets of the brain have made reconstructions of neurons possible at the nanometer scale. However, automatic segmentation sometimes struggles to segment large neurons correctly, requiring human effort to proofread its output. General proofreading involves inspecting large volumes to correct segmentation errors at the pixel level, a visually intensive and time-consuming process. This paper presents the design and implementation of an analytics framework that streamlines proofreading, focusing on connectivity-related errors. We accomplish this with automated likely-error detection and synapse clustering that drives the proofreading effort with highly interactive 3D visualizations. In particular, our strategy centers on proofreading the local circuit of a single cell to ensure a basic level of completeness. We demonstrate our framework’s utility with a user study and report quantitative and subjective feedback from our users. Overall, users find the framework more efficient for proofreading, understanding evolving graphs, and sharing error correction strategies. 

   more » « less
3. Weakly Supervised Deep Nuclei Segmentation using Points Annotation in Histopathology Images

   Qu, H; Wu, P.; Huang, Q.; Yi, J.; Riedlinger, G. M.; De, S.; Metaxas, D. (July 2019, Proceedings of Machine Learning Research)

   Nuclei segmentation is a fundamental task in histopathological image analysis. Typically, such segmentation tasks require significant effort to manually generate pixel-wise annotations for fully supervised training. To alleviate the manual effort, in this paper we propose a novel approach using points only annotation. Two types of coarse labels with complementary information are derived from the points annotation, and are then utilized to train a deep neural network. The fully- connected conditional random field loss is utilized to further refine the model without introducing extra computational complexity during inference. Experimental results on two nuclei segmentation datasets reveal that the proposed method is able to achieve competitive performance compared to the fully supervised counterpart and the state-of-the-art methods while requiring significantly less annotation effort. Our code is publicly available. 

   more » « less
4. PyReconstruct: A fully opensource, collaborative successor to Reconstruct

   https://doi.org/10.1101/2025.04.21.649793  

   Chirillo, Michael A; Falco, Julian N; Musslewhite, Michael D; Lindsey, Larry F; Harris, Kristen M (April 2025, bioRxiv)

   Abstract As the serial section community transitions to volume electron microscopy, tools are needed to balance rapid segmentation efforts with documenting the fine detail of structures that support cell function. New annotation applications should be accessible to users and meet the needs of the neuroscience and connectomics communities while also being useful across other disciplines. Issues not currently addressed by a single, modern annotation application include: 1) built-in curation systems with utilities for expert intervention to provide quality assurance, 2) integrated alignment features that allow for image registration on-the-fly as image flaws are discovered during annotation, 3) simplicity for non-specialists within and beyond the neuroscience community, 5) a system to store experimental meta-data with annotation data in a way that researchers remain masked regarding condition to avoid potential biases, 6) local management of large datasets, 7) fully open-source codebase allowing development of new tools, and more. Here, we present PyReconstruct, a modern successor to the Reconstruct annotation tool. PyReconstruct operates in a field-agnostic manner, runs on all major operating systems, breaks through legacy RAM limitations, features an intuitive and collaborative curation system, and employs a flexible and dynamic approach to image registration. It can be used to analyze, display, and publish experimental or connectomics data. PyReconstruct is suited for generating ground truth to implement in automated segmentation, outcomes of which can be returned to PyReconstruct for proofreading and quality control. Significance statementIn neuroscience, the emerging field of connectomics has produced annotation tools for reconstruction that prioritize circuit connectivity across microns to centimeters and farther. Determining the strength of synapses forming the connections is crucial to understand function and requires quantification of their nanoscale dimensions and subcellular composition. PyReconstruct, successor to the early annotation tool Reconstruct, meets these requirements for synapses and other structures well beyond neuroscience. PyReconstruct lifts many restrictions of legacy Reconstruct and offers a user-friendly interface, integrated curation, dynamic alignment, nanoscale quantification, 3D visualization, and more. Extensive compatibility with third-party software provides access to the expanding tools from the connectomics and imaging communities. 

   more » « less
5. A Continual Learning Framework for Uncertainty-Aware Interactive Image Segmentation

   Zheng, Ervine; Yu, Qi; Li, Rui; Shi, Pengcheng; Haake, Anne (May 2021, Proceedings of the AAAI Conference on Artificial Intelligence)

   null (Ed.)

   Deep learning models have achieved state-of-the-art performance in semantic image segmentation, but the results provided by fully automatic algorithms are not always guaranteed satisfactory to users. Interactive segmentation offers a solution by accepting user annotations on selective areas of the images to refine the segmentation results. However, most existing models only focus on correcting the current image’s misclassified pixels, with no knowledge carried over to other images. In this work, we formulate interactive image segmentation as a continual learning problem and propose a framework to effectively learn from user annotations, aiming to improve the segmentation on both the current image and unseen images in future tasks while avoiding deteriorated performance on previously-seen images. It employs a probabilistic mask to control the neural network’s kernel activation and extract the most suitable features for segmenting images in each task. We also design a task-aware architecture to automatically infer the optimal kernel activation for initial segmentation and subsequent refinement. Interactions with users are guided through multi-source uncertainty estimation so that users can focus on the most important areas to minimize the overall manual annotation effort. Extensive experiments are performed on both medical and natural image datasets to illustrate the proposed framework’s effectiveness on basic segmentation performance, forward knowledge transfer, and backward knowledge transfer. 

   more » « less